Custom MCP Server on Databricks Apps

pillar_id,principle_id,measure_id,alaysis_id,measure_sql_code,measure_sql_description DG,DG-01,DG-01-01,DG-01-01,, DG,DG-01,DG-01-02,DG-01-02,, DG,DG-01,DG-01-03,DG-01-03,, DG,DG-01,DG-01-04,DG-01-04,, DG,DG-01,DG-01-05,DG-01-05,, DG,DG-01,DG-01-06,DG-01-06,, DG,DG-02,DG-02-01,DG-02-01,, DG,DG-02,DG-02-02,DG-02-02,, DG,DG-02,DG-02-03,DG-02-03,, DG,DG-03,DG-03-01,DG-03-01,, DG,DG-03,DG-03-02,DG-03-02,, DG,DG-03,DG-03-03,DG-03-03,, IU,IU-01,IU-01-01,IU-01-01,, IU,IU-01,IU-01-02,IU-01-02,, IU,IU-01,IU-01-03,IU-01-03,, IU,IU-01,IU-01-04,IU-01-04,, IU,IU-02,IU-02-01,IU-02-01,, IU,IU-02,IU-02-02,IU-02-02,, IU,IU-02,IU-02-03,IU-02-03,, IU,IU-03,IU-03-01,IU-03-01,, IU,IU-03,IU-03-02,IU-03-02,, IU,IU-03,IU-03-03,IU-03-03,, IU,IU-03,IU-03-04,IU-03-04,, IU,IU-04,IU-04-01,IU-04-01,, IU,IU-04,IU-04-02,IU-04-02,, IU,IU-04,IU-04-03,IU-04-03,, OE,OE-01,OE-01-01,OE-01-01,, OE,OE-01,OE-01-02,OE-01-02,, OE,OE-01,OE-01-03,OE-01-03,, OE,OE-01,OE-01-04,OE-01-04,, OE,OE-01,OE-01-05,OE-01-05,, OE,OE-01,OE-01-06,OE-01-06,, OE,OE-01,OE-01-07,OE-01-07,, OE,OE-01,OE-01-08,OE-01-08,, OE,OE-01,OE-01-09,OE-01-09,, OE,OE-02,OE-02-01,OE-02-01,, OE,OE-02,OE-02-02,OE-02-02,, OE,OE-02,OE-02-03,OE-02-03,, OE,OE-02,OE-02-04,OE-02-04,, OE,OE-02,OE-02-05,OE-02-05,, OE,OE-02,OE-02-06,OE-02-06,, OE,OE-02,OE-02-07,OE-02-07,, OE,OE-02,OE-02-08,OE-02-08,, OE,OE-02,OE-02-09,OE-02-09,, OE,OE-02,OE-02-10,OE-02-10,, OE,OE-02,OE-02-11,OE-02-11,, OE,OE-03,OE-03-01,OE-03-01,, OE,OE-03,OE-03-02,OE-03-02,, OE,OE-03,OE-03-03,OE-03-03,, OE,OE-03,OE-03-04,OE-03-04,, OE,OE-04,OE-04-01,OE-04-01,, OE,OE-04,OE-04-02,OE-04-02,, SC,SCP-0,SCP-01-01,SCP-01-01,, SC,SCP-0,SCP-01-02,SCP-01-02,, SC,SCP-0,SCP-01-03,SCP-01-03,, SC,SCP-0,SCP-01-04,SCP-01-04,, SC,SCP-0,SCP-01-05,SCP-01-05,, SC,SCP-0,SCP-01-06,SCP-01-06,, SC,SCP-0,SCP-01-07,SCP-01-07,, SC,SCP-0,SCP-01-08,SCP-01-08,, SC,SCP-0,SCP-01-09,SCP-01-09,, SC,SCP-0,SCP-01-10,SCP-01-10,, SC,SCP-0,SCP-01-11,SCP-01-11,, SC,SCP-0,SCP-01-12,SCP-01-12,, SC,SCP-0,SCP-01-13,SCP-01-13,, SC,SCP-0,SCP-02-01,SCP-02-01,, SC,SCP-0,SCP-02-02,SCP-02-02,, SC,SCP-0,SCP-02-03,SCP-02-03,, SC,SCP-0,SCP-02-04,SCP-02-04,, SC,SCP-0,SCP-02-05,SCP-02-05,, SC,SCP-0,SCP-02-06,SCP-02-06,, SC,SCP-0,SCP-02-07,SCP-02-07,, SC,SCP-0,SCP-03-01,SCP-03-01,, SC,SCP-0,SCP-03-02,SCP-03-02,, SC,SCP-0,SCP-03-03,SCP-03-03,, SC,SCP-0,SCP-03-04,SCP-03-04,, SC,SCP-0,SCP-03-05,SCP-03-05,, SC,SCP-0,SCP-03-06,SCP-03-06,, SC,SCP-0,SCP-04-01,SCP-04-01,, SC,SCP-0,SCP-05-01,SCP-05-01,, SC,SCP-0,SCP-06-01,SCP-06-01,, SC,SCP-0,SCP-06-02,SCP-06-02,, SC,SCP-0,SCP-06-03,SCP-06-03,, SC,SCP-0,SCP-06-04,SCP-06-04,, SC,SCP-0,SCP-06-05,SCP-06-05,, SC,SCP-0,SCP-07-01,SCP-07-01,, SC,SCP-0,SCP-07-02,SCP-07-02,, SC,SCP-0,SCP-07-03,SCP-07-03,, SC,SCP-0,SCP-07-04,SCP-07-04,, SC,SCP-0,SCP-07-05,SCP-07-05,, R,R-01,R-01-01,R-01-01,, R,R-01,R-01-02,R-01-02,, R,R-01,R-01-03,R-01-03,, R,R-01,R-01-04,R-01-04,, R,R-01,R-01-05,R-01-05,, R,R-01,R-01-06,R-01-06,, R,R-02,R-02-01,R-02-01,, R,R-02,R-02-02,R-02-02,, R,R-02,R-02-03,R-02-03,, R,R-02,R-02-04,R-02-04,, R,R-02,R-02-05,R-02-05,, R,R-03,R-03-01,R-03-01,, R,R-03,R-03-02,R-03-02,, R,R-04,R-04-01,R-04-01,, R,R-04,R-04-02,R-04-02,, R,R-04,R-04-03,R-04-03,, R,R-04,R-04-04,R-04-04,, R,R-05,R-05-01,R-05-01,, R,R-05,R-05-02,R-05-02,, PE,PE-01,PE-01-01,PE-01-01,, PE,PE-01,PE-01-02,PE-01-02,, PE,PE-02,PE-02-01,PE-02-01,, PE,PE-02,PE-02-02,PE-02-02,, PE,PE-02,PE-02-03,PE-02-03,, PE,PE-02,PE-02-04,PE-02-04,, PE,PE-02,PE-02-05,PE-02-05,, PE,PE-02,PE-02-06,PE-02-06,, PE,PE-02,PE-02-07,PE-02-07,, PE,PE-02,PE-02-08,PE-02-08,, PE,PE-02,PE-02-09,PE-02-09,, PE,PE-02,PE-02-10,PE-02-10,, PE,PE-02,PE-02-11,PE-02-11,, PE,PE-02,PE-02-12,PE-02-12,, PE,PE-02,PE-02-14,PE-02-14,, PE,PE-02,PE-02-15,PE-02-15,, PE,PE-02,PE-02-16,PE-02-16,, PE,PE-03,PE-03-01,PE-03-01,, PE,PE-03,PE-03-02,PE-03-02,, PE,PE-03,PE-03-03,PE-03-03,, PE,PE-04,PE-04-01,PE-04-01,, PE,PE-04,PE-04-02,PE-04-02,, PE,PE-04,PE-04-03,PE-04-03,, CO,CO-01,CO-01-01,CO-01-01A,"SELECT data_source_format AS tables_format, count(data_source_format) AS no_of_tables FROM system.information_schema.tables GROUP BY ALL ORDER BY no_of_tables desc;","This query displays the number of tables registered in the current workspace’s Unity Catalog, grouped by table format (for example, Delta, Iceberg, or Parquet)." CO,CO-01,CO-01-01,CO-01-01B,"SELECT table_type, round(count(table_type)/(select count(*) from system.information_schema.tables) * 100) as percent_of_tables FROM system.information_schema.tables group by ALL HAVING percent_of_tables > 0 ORDER BY percent_of_tables desc","This query shows the percentage breakdown of table types in the current Databricks workspace that are classified as managed tables (for example, Views, Managed, or External)." CO,CO-01,CO-01-02,CO-01-02,"with clusters AS ( SELECT *, ROW_NUMBER() OVER(PARTITION BY workspace_id, cluster_id ORDER BY change_time DESC) as rn FROM system.compute.clusters WHERE cluster_source=""UI"" OR cluster_source=""API"" QUALIFY rn=1 ), job_tasks_exploded AS ( SELECT workspace_id, job_id, EXPLODE(compute_ids) as cluster_id FROM system.lakeflow.job_task_run_timeline WHERE period_start_time >= CURRENT_DATE() - INTERVAL 30 DAY GROUP BY ALL ), all_purpose_cluster_jobs AS ( SELECT t1.*, t2.cluster_name, t2.owned_by FROM job_tasks_exploded t1 INNER JOIN clusters t2 USING (workspace_id, cluster_id) ) SELECT * FROM all_purpose_cluster_jobs LIMIT 10;",This query lists the top ten jobs configured to run on All-Purpose clusters instead of Job clusters within the current Databricks workspace. CO,CO-01,CO-01-03,CO-01-03,"SELECT billing_origin_product, sum(usage_quantity) as dbu FROM system.billing.usage WHERE billing_origin_product in ('SQL','ALL_PURPOSE') AND usage_date >= current_date() - interval 30 days GROUP BY billing_origin_product; ",This query shows the percentage of workloads executed in the last 30 days that used the more cost-effective Databricks SQL Warehouse instead of All-Purpose compute for SQL and BI workloads. CO,CO-01,CO-01-04,CO-01-04,"SELECT cast(regexp_extract(dbr_version, '^(\\d+\\.\\d+)',1) as decimal(3,1)) as dbr_version, count(*) as count FROM system.compute.clusters WHERE NOT contains(dbr_version, 'custom') AND cluster_source not in('PIPELINE','PIPELINE_MAINTENANCE') AND delete_time is null GROUP by 1 ORDER BY dbr_version DESC;","This query displays the number of clusters running on each Databricks Runtime version in the current workspace. Ideally, all clusters should be using the latest runtime version." CO,CO-01,CO-01-05,CO-01-05,, CO,CO-01,CO-01-06,CO-01-06A,"WITH serverless AS ( SELECT sum(usage_quantity) as dbu FROM system.billing.usage u WHERE contains(u.sku_name, 'SERVERLESS') AND u.billing_origin_product in ('ALL_PURPOSE','SQL','JOBS', 'DLT','INTERACTIVE') AND date_diff(day, u.usage_start_time, now()) <28 ), total AS ( SELECT sum(usage_quantity) as dbu FROM system.billing.usage u WHERE u.billing_origin_product in ('ALL_PURPOSE','SQL','JOBS', 'DLT','INTERACTIVE') AND date_diff(day, u.usage_start_time, now()) <28 ) SELECT serverless.dbu * 100 / total.dbu as serverless_dbu_percent FROM serverless CROSS JOIN total;",This SQL query calculates the percentage of overall workloads in the current Databricks workspace that are using Serverless compute. CO,CO-01,CO-01-06,CO-01-06B,"SELECT CASE WHEN t1.sku_name LIKE '%SERVERLESS_SQL%' THEN 'SQL_SERVERLESS' WHEN t1.sku_name LIKE '%ENTERPRISE_SQL_COMPUTE%' THEN 'SQL_CLASSIC' WHEN t1.sku_name LIKE '%SQL_PRO%' THEN 'SQL_PRO' ELSE 'Other' END as sql_sku_name, SUM(t1.usage_quantity * list_prices.pricing.default) as list_cost FROM system.billing.usage t1 INNER JOIN system.billing.list_prices ON t1.cloud = list_prices.cloud AND t1.sku_name = list_prices.sku_name AND t1.usage_start_time >= list_prices.price_start_time AND (t1.usage_end_time <= list_prices.price_end_time OR list_prices.price_end_time IS NULL) WHERE t1.sku_name LIKE '%SQL%' AND t1.usage_date >= current_date() - interval 30 days GROUP BY ALL;","This SQL query shows the cost distribution of SQL Warehouses by type, highlighting the proportion of Serverless usage compared with Classic compute within the current workspace" CO,CO-01,CO-01-07,CO-01-07,, CO,CO-01,CO-01-08,CO-01-08,"WITH per_cluster_daily AS ( SELECT cluster_id, DATE_TRUNC('DAY', start_time) AS day, AVG(cpu_user_percent + cpu_system_percent) AS avg_cpu_usage_percent, AVG(mem_used_percent) AS avg_memory_usage_percent FROM system.compute.node_timeline WHERE start_time >= CURRENT_DATE - INTERVAL 28 DAYS GROUP BY cluster_id, DATE_TRUNC('DAY', start_time) ) SELECT percentile(avg_cpu_usage_percent, 0.75) as cpu_usage_percent_p75, percentile(avg_memory_usage_percent, 0.75) as memory_usage_percent_p75 FROM per_cluster_daily GROUP BY ALL;","This SQL query reports the 75th percentile (P75) CPU and memory utilisation for compute resources configured in the current Databricks workspace. " CO,CO-01,CO-01-09,CO-01-09,, CO,CO-02,CO-02-01,CO-02-01,"WITH autoscaling_count AS ( SELECT count(*) as autoscaling_count FROM system.compute.clusters WHERE max_autoscale_workers IS NOT NULL AND delete_time IS NULL ), total_clusters_count AS ( SELECT count(*) as total_clusters_count FROM system.compute.clusters WHERE delete_time IS NULL ) SELECT autoscaling_count.autoscaling_count * 100 / total_clusters_count.total_clusters_count as autoscaling_percent FROM total_clusters_count CROSS JOIN autoscaling_count;",This query shows the percentage of clusters with auto-scaling enabled in the current Databricks workspace. CO,CO-02,CO-02-02,CO-02-02,"SELECT percentile(c.auto_termination_minutes, 0.75) as p_75_auto_termination_minutes, max(c.auto_termination_minutes) as max_auto_termination_minutes, count_if(c.auto_termination_minutes is null) as count_clusters_without_autoterminations, count_if(c.auto_termination_minutes is not null) as count_clusters_with_autoterminations, (count_clusters_without_autoterminations*100)/count(*) as percent_clusters_without_autoterminations FROM system.compute.clusters c WHERE c.cluster_source in ('UI','API') AND c.delete_time IS NULL;",This query reports the number of Databricks usage table reads over the past 30 days within the workspace — a strong indicator of whether cost observability dashboards and similar monitoring tools are in use. CO,CO-02,CO-02-03,CO-02-03,, CO,CO-03,CO-03-01,CO-03-01,"SELECT count(*) as usage_read FROM system.access.audit WHERE service_name = 'unityCatalog' AND action_name = 'getTable' AND request_params.full_name_arg = 'system.billing.usage' AND user_identity.email != 'System-User' AND (date_diff(day, event_date, current_date()) <= 30);","This query reports the number of Databricks usage table reads over the past 30 days within the workspace — a strong indicator of whether cost observability dashboards and similar monitoring tools are in use. " CO,CO-03,CO-03-02,CO-03-02a,"SELECT array_size(map_entries(tags)) as number_of_tags, count(*) as count FROM system.compute.clusters WHERE tags.ResourceClass IS NULL AND delete_time IS NULL GROUP BY number_of_tags ORDER BY count DESC, number_of_tags DESC;","This query counts the number of compute instances configured in the workspace, grouped by the number of tags applied. Tagging is a best practice for cost observability and governance." CO,CO-03,CO-03-02,CO-03-02b,"WITH tag_counts AS ( SELECT explode(map_keys(tags)) as tag, count(*) as count FROM system.compute.clusters GROUP BY 1 ), cluster_count AS (SELECT count(*) as count FROM system.compute.clusters) SELECT tag_counts.tag, sum(tag_counts.count) / any_value(cluster_count.count) *100 as percent FROM tag_counts CROSS JOIN cluster_count GROUP BY tag_counts.tag ORDER BY percent DESC;","This query shows the percentage breakdown of tags used within the workspace, helping identify gaps or inconsistencies in tagging policies." CO,CO-03,CO-03-03,CO-03-03,, CO,CO-03,CO-03-04,CO-03-04,, CO,CO-03,CO-03-05,CO-03-05,, CO,CO-04,CO-04-01,CO-04-01,, CO,CO-04,CO-04-02,CO-04-02,,